Article Synthetic Lethality between DNA Polymerase Epsilon and RTEL1 in Metazoan DNA Replication Graphical Abstract Authors Roberto Bellelli, Jillian Youds, Valerie Borel, Jennifer Svendsen, Visnja Pavicic-Kaltenbrunner, Simon J. Boulton Correspondence [email protected] In Brief Bellelli et al. report that RTEL1 deficiency is synthetic lethal with the loss of pole-4 in C. elegans/hypomorphy of Pol epsilon. An analysis of replication dynamics in Rtel1À/À Pole4À/À mouse cells revealed a combination of dysfunctional fork progression and defective origin activation, which cooperatively drive incomplete genomic replication and genetic instability. Highlights d rtel-1 is synthetic lethal with the loss of DNA polymerase epsilon in C. elegans d rtel-1; pole-4 double mutants accumulate Rad51 and RPA foci and fail to replicate d Impaired DNA replication and genome instability in Rtel1 Pole4 knockout mouse cells d Rtel1 Pole4 double knockout mouse cells exhibit fork asymmetry and defective origin activation Bellelli et al., 2020, Cell Reports 31, 107675 May 26, 2020 ª 2020 The Author(s). https://doi.org/10.1016/j.celrep.2020.107675 ll ll OPEN ACCESS Article Synthetic Lethality between DNA Polymerase Epsilon and RTEL1 in Metazoan DNA Replication Roberto Bellelli,1,2,3 Jillian Youds,1,2 Valerie Borel,1 Jennifer Svendsen,1 Visnja Pavicic-Kaltenbrunner,1 and Simon J. Boulton1,4,* 1The Francis Crick Institute, 1 Midland Road, NW1 1AT London, UK 2These authors contributed equally 3Present address: Center for Cancer Cell and Molecular Biology, Barts Cancer Institute, Queen Mary University of London, Charterhouse Square, Barbican, EC1M 6BE London, UK 4Lead Contact *Correspondence: [email protected] https://doi.org/10.1016/j.celrep.2020.107675 SUMMARY Genome stability requires coordination of DNA replication origin activation and replication fork progression. RTEL1 is a regulator of homologous recombination (HR) implicated in meiotic cross-over control and DNA repair in C. elegans. Through a genome-wide synthetic lethal screen, we uncovered an essential genetic inter- action between RTEL1 and DNA polymerase (Pol) epsilon. Loss of POLE4, an accessory subunit of Pol epsilon, has no overt phenotype in worms. In contrast, the combined loss of POLE-4 and RTEL-1 results in embryonic lethality, accumulation of HR intermediates, genome instability, and cessation of DNA replica- tion. Similarly, loss of Rtel1 in Pole4À/À mouse cells inhibits cellular proliferation, which is associated with persistent HR intermediates and incomplete DNA replication. We propose that RTEL1 facilitates genome- wide fork progression through its ability to metabolize DNA secondary structures that form during DNA repli- cation. Loss of this function becomes incompatible with cell survival under conditions of reduced origin acti- vation, such as Pol epsilon hypomorphy. INTRODUCTION size and viability, sensitivity to DNA damaging agents, and elevated meiotic recombination. Biochemical studies show DNA replication origins are established at thousands of sites that RTEL1 can efficiently disassemble D-loop recombination in- throughout the genome through a combination of structural termediates, suggesting that RTEL-1 might disassemble these and functional chromatin determinants that promote loading of intermediates to promote non-crossover repair, likely through inactive MCM2-7 double hexamers around DNA replication ori- synthesis-dependent strand annealing. Genetic analysis also re- gins (Fragkos et al., 2015). At the G1-S transition, DDK- and vealed that rtel-1 is synthetic lethal when combined with muta- CDK-dependent phosphorylation events drive the formation tions in dog-1/FANCJ, mus-81, him-6/BLM, and rcq-5, all of and activation of the CMG (CDC45-MCM2-7-GINS), the eukary- which are homologs of genes involved in human genetic dis- otic replicative helicase, followed by establishment of two sym- eases and the maintenance of genome stability at replication metric replication forks that initiate DNA synthesis (Bell and La- forks (Barber et al., 2008). Indeed, mutants of these genes bib, 2016). The bulk of DNA replication at active replication when combined with rtel-1 displayed persistent RAD-51 foci in forks is performed by the conserved DNA polymerase com- the germline and embryonic lethality, indicating that RTEL-1 is plexes Pol delta and Pol epsilon, which act on the lagging and essential in their absence. However, where and when RTEL1 ac- leading strand, respectively (Burgers and Kunkel, 2017). The tivity is essential in nematodes remained to be established. Sub- latter is also considered to be a stable component of the repli- sequent studies showed that RTEL1 facilitates efficient telomere some and is required for efficient CMG formation in budding and genome-wide replication in vertebrates (Vannier et al., yeast (Bell and Labib, 2016). DNA replication in metazoans 2013), but its precise function during DNA replication remains also requires the function of several helicases and replisome- unclear. associated factors to prevent inappropriate transactions at the To gain an improved understanding of the role of RTEL-1 in replication fork and to avert persistent fork stalling events maintaining genome stability, we conducted a genome-wide (Leo´ n-Ortiz et al., 2014; Dungrawala et al., 2015). RNAi screen to identify genes that, when knocked down with The helicase RTEL1 was identified as the first metazoan anti- RNAi in the rtel-1 mutant background, cause synthetic lethality, recombinase, which facilitates DNA repair and regulates cross but not in the wild-type. This genetic screen identified multiple over formation in meiosis in C. elegans (Barber et al., 2008; genes involved in DNA replication, such as TOPBP1, GINS com- Youds et al., 2010). rtel-1 mutant worms display reduced brood plex subunits PSF2 and PSF3, RFC-1, FEN-1, and CDT1, as well Cell Reports 31, 107675, May 26, 2020 ª 2020 The Author(s). 1 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). ll OPEN ACCESS Article as three components of DNA Pol epsilon, which we chose to extend to DNA Pol delta (Figure S1C). pole-3 is predicted to further investigate. Strains lacking the non-essential subunit of be a non-essential component of the complex, and therefore, Pol epsilon, pole-4, exhibit no overt phenotype under normal RNAi of this gene had a less dramatic effect on rtel-1 mutants or DNA damaging conditions. Strikingly, however, the pole-4; than pole-1 or pole-2 RNAi (Figure 1B). We proceeded to rtel-1 double mutant is 100% synthetic embryonic lethal and pre- examine the germlines of these animals for defects that could sents with persistent homologous recombination (HR) intermedi- explain the source of the observed synthetic lethality. After ates, extensive genome instability, and cessation of DNA RNAi for pole-1 or pole-2, rtel-1 mutants exhibited increased replication. We proceed to show that this synthetic lethal interac- levels of RAD-51 foci in the mitotic zone of the germline, which tion is conserved in mammalian cells. A combined loss of RTEL1 is the only region of active DNA replication in the C. elegans and POLE4 in primary mouse cells also inhibits cellular prolifera- germline. In contrast, RAD-51 foci were not observed in tion and results in extensive genetic instability. Molecular anal- N2(wild-type) animals fed with the same RNAi, suggesting ysis of DNA replication dynamics in Rtel1-Pole4 double that replication defects might underlie the lethality in rtel-1; knockout cells revealed a combination of dysfunctional fork pro- pole-1(RNAi) and rtel-1; pole-2(RNAi) animals (Figure 1C). gression and origin activation, which leads to fork stalling and genome under-replication. rtel-1 Is Synthetic Lethal with pole-4 in C. elegans Our data, although pointing to conserved functions in meta- To confirm our observations with RNAi, we obtained a genetic zoans for RTEL1 in replication fork progression and POLE4 in deletion of pole-4, tm4613, which removed the majority of the maintaining Pol epsilon complex stability, reveal an un-appreci- coding region of the gene, apart from the first 35 nucleotides ated interplay between replication origin activation and fork pro- of exon 1; this likely represents a bona fide null allele (Figure S2A). gression required for genome-wide DNA replication and the pole-4(tm4613) mutants appeared superficially wild type, and maintenance of genome stability. the loss of pole-4 did not result in any significant loss of viability. rtel-1 mutants show greater than 90% viability as previously RESULTS described (Barber et al., 2008). In contrast, rtel-1; pole-4 double mutant animals displayed a dramatic synthetic embryonic lethal rtel-1 Is Synthetic Lethal with Members of the DNA phenotype (Figure 2A), wherein none of the progeny were viable. Polymerase Epsilon Complex in C. elegans This was associated with a dramatic reduction in DNA replica- RTEL1 is dispensable for viability in the nematode C. elegans,an tion, as observed by attenuated incorporation of Cy3-dUTP in organism amenable to genetic manipulation and a potent system double mutant worms (Figure 2B). to identify synthetic interactions in vivo. To further interrogate the Because DNA Pol epsilon has been implicated in activation of functions of RTEL1 in genome stability in C. elegans, we per- the intra-S-phase checkpoint, we questioned whether the formed a genome-wide RNAi screening in N2(wild type) and absence of pole-4 might also inactivate the replication check- rtel-1 mutant worms (Kamath et al., 2003) by using a library of point in worms (Navas et al., 1995). However, in contrast to 16,256 genes (Figure S1A). Following secondary screens to this hypothesis, DAPI staining of the germlines of rtel-1; pole-4 confirm our initial hits, we identified a number of genes with es- double mutants showed enlarged mitotic nuclei and fewer nuclei tablished roles in DNA replication whose RNAi caused lethality in in the mitotic zone than wild-type animals or either single mutant, the rtel-1 mutant but not in an N2(wild-type) strain (Figure S1B).